> For the complete documentation index, see [llms.txt](https://docs.buildings.ability.abb/collection/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.buildings.ability.abb/collection/academy/building-intelligence-ai-and-machine-learning-in-buildingpro-suites/machine-learning/machine-learning-fundamentals-for-buildingpro-suites.md). # Machine Learning Fundamentals for BuildingPro Suites ## Introduction to Machine Learning Machine Learning is a subset of artificial intelligence that focuses on systems which learn patterns from data and use those patterns to make predictions, classifications, or other useful inferences. Instead of relying only on explicitly written rules, Machine Learning models improve their performance by being trained on examples. In the context of smart buildings, this is important because buildings generate large amounts of operational data from many different systems. ABB Ability™ BuildingPro Suites is positioned as a unified cloud intelligence platform that connects building automation, Heating, Ventilation, and Air Conditioning (HVAC), energy, Information Technology (IT), Operational Technology (OT), and other systems to transform siloed building data into actionable insights, dashboards, predictive tools, and intelligence. This creates a strong basis for Machine Learning applications. This page explains what Machine Learning is, how it differs from general artificial intelligence, how it works in principle, and why it is relevant for BuildingPro Suites. ### What is Machine Learning? Machine Learning is the field of artificial intelligence concerned with algorithms that learn from data. A Machine Learning model is not programmed with every decision rule in advance. Instead, it is trained on examples so that it can recognize patterns and apply them to new data. The central idea is simple. If a model learns from data that is representative of the task it will later face in practice, it can generalize from that experience and produce useful outputs when deployed. These outputs may be predictions, anomaly signals, classifications, rankings, or recommendations. In this sense, Machine Learning is less about fixed instructions and more about learned behavior from data. ### Machine Learning and Artificial Intelligence Machine Learning and artificial intelligence are closely related, but they are not the same thing. Artificial intelligence is the broader term. It includes any system that can use information to make decisions or support actions without continuous human intervention. Some artificial intelligence systems are fully rule-based. These systems behave according to explicit logic defined by people. Machine Learning is a subset of artificial intelligence. Here, the logic is not defined completely by hand. Instead, the system learns patterns from examples. A simple building example helps explain the difference. A rule-based system might work like this: * if room temperature is below a defined threshold, start heating * if room temperature is above a defined threshold, start cooling This is useful and often necessary. But it is still based on fixed rules. A Machine Learning system works differently. Instead of relying only on manually defined thresholds, it can learn from historical behavior, equipment states, time patterns, and other context to identify expected behavior, estimate demand, or detect unusual conditions. In practice, smart building solutions often combine both approaches: * rules for deterministic behavior * analytics for visibility * Machine Learning for learned prediction and pattern recognition ### How Machine Learning works At a high level, Machine Learning works by learning a relationship between input data and a desired output. The process usually starts with data. In a building context, this data can include measured values, device states, alarms, events, asset metadata, schedules, or contextual information such as time and location. During training, the model is exposed to this data and adjusts its internal parameters so that its outputs become more useful for the intended task. The goal is not simply to perform well on training data. The real goal is generalization. This means the model should also perform well on new data that it has not seen before. When the model is later used in a real application, it applies what it has learned to incoming data. This step is often called inference. The practical value of Machine Learning therefore depends on three things: * relevant and reliable data * a model suitable for the problem * useful generalization to real building scenarios ### Why Machine Learning is useful in smart buildings Smart buildings contain many systems, many assets, and many data points. As the number of connected devices and systems grows, it becomes harder to understand all relevant interactions through manual review alone. Machine Learning is useful in this environment because it can help reveal patterns that are difficult to define with fixed logic. It is especially relevant when users need to answer questions such as: * What is likely to happen next? * What behavior looks unusual? * Which asset or issue should be investigated first? * Which trends matter across a portfolio? This aligns well with the public positioning of BuildingPro Suites, which includes predictive tools, smarter data-driven decisions, demand prediction, anomaly discovery, and proactive AI \[1]. ### Main types of Machine Learning Machine Learning methods are often grouped into three broad categories. #### Supervised learning In supervised learning, a model is trained using examples where the desired output is already known. The goal is to learn a mapping from input to output. This approach is commonly used for tasks such as: * prediction of future values * classification of events or conditions * scoring or ranking problems #### Unsupervised learning In unsupervised learning, the model works with unlabeled data and tries to identify structure or patterns on its own. This is useful for tasks such as: * clustering similar assets or behaviors * finding correlations * detecting unusual patterns * reducing data complexity #### Reinforcement learning In reinforcement learning, a system learns through interaction, feedback, and reward. Instead of predicting one correct answer, it learns which actions are better over time. This is especially relevant for sequential decision-making and optimization problems, and it is discussed in a separate academy section. ### Deep learning and modern Machine Learning Over the last years, deep learning has become one of the most important approaches in Machine Learning. Deep learning uses large artificial neural networks that can learn complex patterns directly from data. This is one reason why modern AI systems have advanced so quickly in areas such as language, images, pattern recognition, and generative AI. Large Language Models (LLMs) and other foundation-model-based systems are part of this broader development. This does not make earlier Machine Learning concepts irrelevant. It means the field has expanded. Many modern AI systems still rely on the same core idea: learning patterns from data and using them to generalize to new situations. For BuildingPro Suites, this is an important perspective. Whether an application uses traditional Machine Learning, deep learning, or newer model types, the value still depends heavily on connected, structured, and meaningful building data. ### Machine Learning in BuildingPro Suites BuildingPro Suites provides several important foundations that support Machine Learning applications. It connects data from building automation, HVAC, energy, IT, OT, and third-party systems into one platform \[1]. It also provides structure through tags, custom asset templates, asset instances, nested assets, points, and attributes \[2]. These capabilities help define what the data represents and how different objects relate to one another. This matters because Machine Learning models are more useful when they operate on data that is not only available, but also well organized and meaningful. A model output becomes far more actionable when it can be linked to a known asset, building area, hierarchy, or workflow. BuildingPro Suites also supports integrations, APIs, and app-based extensions in a scalable architecture \[2]. This is important because Machine Learning needs more than models. It also needs a way to access data, run services, and deliver outputs into operational workflows. ### Conclusion Machine Learning is a method for building systems that learn from data instead of relying only on fixed instructions. It is a core part of modern artificial intelligence and is widely used to support prediction, anomaly detection, classification, and prioritization. In smart buildings, Machine Learning is particularly valuable because it helps turn large amounts of connected operational data into more useful insight. BuildingPro Suites provides important foundations for this through data integration, semantic structure, scalable architecture, and predictive intelligence positioning. For this reason, Machine Learning is not just a technical topic in BuildingPro Suites. It is part of the broader goal of transforming building data into practical intelligence for buildings and portfolios. --- # Agent Instructions This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com. ## Querying This Documentation If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question. Perform an HTTP GET request on the current page URL with the `ask` query parameter, and the optional `goal` query parameter: ``` GET https://docs.buildings.ability.abb/collection/academy/building-intelligence-ai-and-machine-learning-in-buildingpro-suites/machine-learning/machine-learning-fundamentals-for-buildingpro-suites.md?ask=&goal= ``` `ask` is the immediate question: it should be specific, self-contained, and written in natural language. `goal` is optional and describes the broader end goal you are ultimately trying to accomplish on behalf of the user. GitBook uses it to tailor the answer towards what is most useful for that goal. The response will contain a direct answer to the question and relevant excerpts and sources from the documentation. Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.